Epistemic status: This is a mechanism-level hypothesis motivated by personal experience, not a rigorous causal account of religion, secularization, or value change in China. I am not claiming that “Chinese people have no faith,” nor that religious decline necessarily produces materialism, deference to power, or moral decline. My narrower question is this: when a society rapidly weakens older religious, traditional, and communal structures, what continues to carry the values that once constrained wealth, power, and competition? And what happens if old carriers disappear faster than new ones can form?
I am not a Christian. Strictly speaking, I do not believe in a personal God in the traditional sense. I find it difficult to believe that behind the universe there is a supernatural person who thinks, becomes angry, shows favoritism, rewards, and punishes in something like the way human beings do. If I had to describe my own position, I would say I am closer to believing in some impersonal ultimate order. But this is only part of what motivated me to think about the problem. It is not a premise on which the argument depends.
What made me reconsider religion was a much simpler question: is a person’s price in society the same thing as that person’s value?
Jobs can disappear. Income can fall. Relationships can end. A person can make a serious effort and still fail. If salary, education, occupation, property, social status, and recognition from others form the final system by which a life is evaluated, then it is easy for “I lost this time” to become “my life is a failure,” and eventually “I myself have little value.”
I need to reject that inference.
That was what gradually made me realize that religion historically addressed far more than the question of who created the universe. It also offered a second system of evaluation. When the ordinary world tells someone, “You lost,” religion can at least make it possible to answer: that does not mean your existence itself has lost.
By measures of formal religious affiliation, China is indeed an unusually nonreligious society. Pew’s synthesis of several Chinese surveys reports that in the 2018 Chinese General Social Survey, about 90% of adults selected no religious affiliation, while only about 34% described themselves as atheists in the 2018 World Values Survey. Belief in Buddhas and bodhisattvas, incense burning, ancestor rites, feng shui, and other practices with religious or supernatural content also continue to exist. “No formal religious affiliation” is therefore clearly not the same as “believing in nothing.” (Pew Research Center)
So the question that interests me is not really, “Why do Chinese people not believe in God?”
It is something closer to this: why can a society that still has family ethics, traditional customs, folk beliefs, and a substantial moral vocabulary nevertheless leave people with the strong impression that money, office, education, property, rank, and power occupy too much of the space in which human value is judged?
That impression is not itself social-scientific evidence. It is only the starting point of the question.
The first distinction I need is between philosophical materialism and a different phenomenon. Philosophical materialism concerns whether reality is fundamentally physical. A committed atheist can perfectly well believe that all people possess equal dignity, that wealth does not determine personhood, that power does not generate truth, and that the weak deserve protection.
That is not what concerns me.
What concerns me is a process in which wealth and human worth become conflated. I will call this wealth-as-worth conflation, rather than “worship” or “fetishism,” because both of those words import theoretical associations that are not necessary for the argument.
It might begin with a simple observation:
“He earns less money.”
Then become:
“He is not doing very well.”
Then:
“He is not very capable.”
And eventually:
“He is not someone worth respecting.”
At that point, wealth has crossed a boundary. It has ceased to function only as a resource in the economic sphere and has entered the sphere of personhood.
This problem is not new. Michael Walzer’s account of “complex equality” in Spheres of Justice develops a closely related idea: an advantage in one social sphere should not automatically be convertible into domination in another. Economic purchasing power, for example, should not automatically purchase political domination. (Stanford Encyclopedia of Philosophy)
Walzer therefore gives us one important question: how do we prevent one social good from becoming dominant and invading domains that should be governed by different principles?
I want to push the question one step further: who actually holds those boundaries?
And if holding the boundary is costly, why should we expect it to survive sustained competition?
Wealth is not the only advantage that can spill across domains. Political office gives someone administrative authority, but administrative authority, epistemic authority, and moral authority are not the same thing. “Who has the right to decide?”, “Who is more likely to know what is true?”, and “Who has the authority to define what is right?” are three separate questions.
If someone who possesses administrative authority is therefore assumed to understand the facts better, and then is further assumed to possess greater moral legitimacy, power has also crossed domains.
So what I have informally thought of as “worship of wealth” or “worship of power” is better described as a form of domain expansion: one kind of advantage begins to determine value in domains where it originally had no legitimate authority.
Why does this happen so easily? I suspect measurability is one important mechanism.
Modern large-scale societies have to allocate resources. Firms have to hire. Schools have to select. Governments have to evaluate. Markets have to price. Large-scale coordination among strangers inevitably creates demand for metrics.
Income is easy to measure.
Exam scores are easy to measure.
Publication counts are easy to measure.
Rank is easy to measure.
Property is easy to measure.
Clicks are easy to measure.
But many of the things we actually care about are much harder to measure.
Is someone trustworthy?
Kind?
Willing to take responsibility when nobody is watching?
Capable of restraining themselves from abusing power?
Willing to protect a stranger who can offer them nothing in return?
These are much harder to compress into stable numerical indicators.
Once competition is organized around metrics, the metrics acquire real power. Once they acquire power, it becomes easy to mistake them for value itself.
This mechanism is not new either. It belongs to the same family of problems as Goodhart’s law and Campbell’s law: once a measure becomes a high-stakes target, people adapt their behavior around the measure, and its relationship with the thing it originally tracked may deteriorate. One formulation discussed by the Bank of England is that a useful statistical relationship may break down once it is turned into a policy or regulatory target, because agents adapt to it. (Bank of England)
So I am not claiming to have rediscovered the fact that metrics can be gamed, nor the idea that money should not be able to buy everything.
What I want to connect are three further questions.
First, what kinds of social structures can actually prevent one kind of advantage from spreading into other domains?
Second, if these structures can be dismantled quickly but rebuilt only slowly, does this create a period of unusual value fragility?
Third, when competitive pressure rises further, especially in the presence of powerful optimizers such as firms, states, and eventually AI systems, which value boundaries can actually survive?
To discuss the first question, I need something narrower than what I previously called a “final veto.” “Final” sounds like an unaccountable supreme authority, which is almost the opposite of what I mean.
A better term is a bounded veto.
A bounded veto is not an authority entitled to decide everything. It is a principle or institution capable of blocking a particular kind of cross-domain conversion.
Money should not directly buy human worth.
Office should not directly create truth.
Majority preference should not, merely by being majority preference, erase the basic rights of a minority.
It is a veto because it can still say “no” when crossing the boundary would be materially profitable. It is bounded because the veto itself remains constrained by other institutions and principles.
A religious institution that claims legitimate authority over politics, science, morality, family life, and private life all at once would therefore not be a successful example of this idea. It would be another case of one sphere colonizing all the others.
Religion has historically been only one possible carrier of these boundaries.
If the idea of a “value carrier” is to be testable rather than a concept that can explain anything after the fact, I think it should be assessed along at least four dimensions.
The first is independence. A body that depends entirely on the institution it is supposed to constrain will have difficulty constraining it for long.
The second is cost-bearing capacity. A veto matters only when saying no actually risks resources, status, or security.
The third is persistence. A society cannot depend indefinitely on a handful of moral heroes. A carrier needs to survive personnel turnover and generational change.
The fourth is generalizability. If we are talking about public values, a carrier cannot protect only family members, co-religionists, colleagues, or insiders. It needs some capacity to extend protection to strangers.
These four dimensions are not a pass-or-fail test. They are dimensions on which real institutions can be stronger or weaker.
A constitutional court with secure tenure, relatively independent funding, and standing for ordinary people to bring rights claims might score reasonably well on independence and generalizability. But if executive actors can simply ignore its decisions, its ability to bear conflict costs may be weak.
A corporate ethics board may be staffed by serious and conscientious people, yet still be weakly independent if the executives it reviews can remove its members at will.
A religious community may persist for centuries and may be unusually capable of absorbing costs, while still performing poorly on generalizability if it recognizes the dignity of insiders but not outsiders.
Religion therefore does not automatically solve the problem.
Societies with strong religious traditions can still display intense wealth competition, status competition, and the sacralization of political authority. Prosperity theology can reinterpret wealth itself as evidence of divine blessing. Religious institutions can also be captured by political power.
The reverse case matters just as much. Highly secular societies can develop robust human-rights norms, legal institutions, civil society, and public trust. Religion is neither a necessary nor a sufficient condition.
My narrower claim is only that the disappearance of old carriers does not automatically create new ones, and the speed at which the two processes occur may differ.
Chinese history itself makes any simple story about an “absence of transcendence” difficult to sustain.
Traditional China had the Dao, the Mandate of Heaven, traditions of remonstrance, and ideals such as sacrificing one’s life for righteousness. These ideas are not equivalent to modern human-rights institutions, but they clearly show that Chinese moral thought was not reducible to obedience, wealth, or immediate material interest. A scholar-official could regard himself as accountable to standards higher than the ruler’s immediate will.
The more interesting question is therefore not why China never had higher-order values.
It is why older higher-order values did not fully translate into modern, stable, relatively independent public carriers.
Family is another difficult counterexample. Many parents are willing to bear enormous costs for their children. Friendship can involve loyalty. Family life contains responsibility and sacrifice. Obviously, not everything has been priced by the market.
But family also exposes another limitation: a value does not become a public value merely because it is non-market.
“I would sacrifice almost anything for my child” does not automatically imply “I recognize the equal and non-negotiable dignity of a stranger’s child.”
Indeed, because parents care deeply about their children, family commitments can intensify competition over housing, school districts, jobs, and educational opportunities.
Family values can resist markets, and they can also become engines of market competition.
That is why generalizability matters.
I do not want to explain modern China by appealing to some fixed national character. China underwent an extraordinary transformation over the twentieth century: war, revolution, state reconstruction, changes in local and religious communities, followed by rapid industrialization, urbanization, marketization, educational expansion, and large-scale migration. Formal religious identification is low, while traditional religious practices have not disappeared entirely. (Pew Research Center)
Marketization, inequality, educational filtering, and resource scarcity may all be more direct explanations for why money and rank carry so much social weight.
My hypothesis adds only one further question:
When competition intensifies rapidly, who prevents competitive logic from expanding into every domain?
This also needs to be a hypothesis that could fail.
If competitive intensity were roughly held constant, my view predicts that societies with more independent, more cost-bearing, more persistent, and more generalizable value carriers should be better able to prevent wealth, office, and power from spreading across domains.
Mainland China and Taiwan might be one comparative case, but certainly not a natural experiment. Other post-socialist societies, different East Asian societies, and highly religious and highly secular societies could provide additional comparisons.
The real empirical task would be to control, as far as possible, for income, inequality, educational competition, institutional quality, and market pressure, and then ask whether what I am calling carrier strength still tracks outcomes such as these:
Do people more readily interpret economic success as evidence of superior human worth?
Does administrative authority more readily become epistemic authority?
Are people more willing to tolerate violations of basic rights for the sake of economic gain?
Is the basic dignity of strangers protected consistently across class and group boundaries?
“Carrier strength” cannot remain an undefined hidden variable either. Empirical work could at least use proxies such as judicial independence, civil-society autonomy, organizational independence, generalized trust, and survey measures concerning rights and authority.
If, after controlling for the more obvious factors, carrier strength has little stable relationship with these outcomes, I would lower my confidence in the hypothesis.
If new secular carriers usually arise rapidly and spontaneously after old value structures collapse, then the timing claim I am emphasizing would also become much less important.
China, in other words, is a motivating case for me, not evidence that the theory has already been established.
This question also has an application to discussions of value erosion in effective altruism. The EA Forum describes value erosion as a class of competitive dynamics in which states, firms, or autonomous AI systems with worse values may expand because they are more competitive, potentially spreading or locking in those values. (EA Forum)
I do not think secularization in modern societies and future AI competition are the same phenomenon. The relevant evidence is very different. But they share an abstract question:
If protecting a value reduces short-term competitive performance, what protects that value?
This is an application of the framework, not where the argument began.
When stronger optimizers appear, whether firms, states, or AI systems, we can ask the same questions. Is the institution protecting a value independent of the optimizer it is supposed to constrain? Can it actually bear the cost of saying no? Can it persist across time? Does it protect a general value, or merely the interests of insiders?
This framework does not automatically produce a policy answer. It may turn out that almost no real institution performs well on all four dimensions. That itself would be informative, because it distinguishes between a value that exists on paper and a value that has a social carrier capable of surviving competitive pressure.
This is ultimately why I have come to understand religion differently without becoming a traditional religious believer.
What now interests me most about religion is no longer whether it can explain the origin of the universe, lightning, or disease. Science is plainly better suited to those questions.
What interests me is why some religious traditions can give certain values a status that is not reducible to current social prices.
A poor person can, in principle, stand before God without being worth less than a rich person.
A king does not possess ultimate truth merely because he is king.
Wealth cannot automatically purchase salvation.
Office does not automatically determine human worth.
Believers can betray these principles. Churches can betray them. Religion itself can be captured by politics and wealth.
So I am not arguing that religion is inherently kinder.
I am saying that religion was one of the historical technologies humans used to answer a difficult question:
How can some values remain stable outside the immediate logic of material advantage?
Continued secularization in Western societies is not necessarily a problem. Highly secular societies have already shown that people can reject belief in a personal God while maintaining morality, human rights, dignity, and public trust.
I am therefore not writing from across the Pacific to tell anyone that they should return to church.
The question I want to ask is narrower.
If older religious carriers continue to weaken, what structures are replacing the functions they once performed?
When someone loses a job, what prevents “the market has less use for me right now” from becoming “I have less value as a person”?
When someone accumulates enormous wealth, what prevents wealth from being converted into moral superiority?
When an institution possesses administrative power, what prevents that power from being converted into epistemic and moral authority?
When a majority benefits from something, what continues to preserve boundaries that should not be traded away for minority groups?
If stable secular answers to these questions already exist, then religious decline may not matter very much.
But if they do not, what remains is not necessarily a purely rational vacuum.
The easiest things to occupy that space are often the things that are easiest to calculate, compare, reward, and optimize:
money,
rank,
office,
attention,
efficiency,
power.
I do not know whether a personal God exists. My own beliefs are closer to some impersonal ultimate order. But even if that belief is entirely mistaken, the problem in this essay remains.
What should not be priced?
What should not become correct merely because it increases efficiency?
What should not be purchasable by wealth?
What should not be definable by power?
What are we willing to protect even when doing so reduces short-term competitive performance?
I no longer think a society necessarily needs to recover God.
I do think it needs values and institutions capable of holding boundaries between domains, while themselves remaining bounded and constrained.
Ceasing to kneel before old gods may indeed be a form of liberation.
But secularization does not automatically complete the rest of the work.
The harder question is this:
After the old gods are gone, have we built anything strong enough, and constrained enough, that we no longer need to kneel before wealth, power, and rank?
